Multicale Analysis of Plane Arrangements (MAPA)

A comprehensive solution to the problem of hybrid linear modeling, that is, to model and segment data using an arrangement of affine subspaces. It only requires the coordinates of the data as input and can automatically determine the optimal model parameters.

Publications

• Multiscale Geometric and Spectral Analysis of Plane Arrangements, G. Chen and M. Maggioni, to appear in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Colorado Springs, CO, June 2011 [PDF]
• Multiscale Geometric Methods for Data Sets III: Multiple Planes, G. Chen and M. Maggioni, in preparation, 2011

Matlab Codes

• MAPA
• Other algorithms (compared with in the CVPR paper):
• ALC
• GPCA-voting
• LBF
• LRR
• SCC
• SSC

Supplemental Data

• Artificial data: generated using 'generate_samples.m' from the GPCA-voting package
• Real data

Contact Info

• Mauro Maggioni, mauro (at) math.duke.edu
• Guangliang Chen, guangliang.chen (at) sjsu.edu

Acknowledgement

• The research described here was supported by NSF DMS 0650413, IIS 0803293, and ONR N00014-07-1-0625.

Last updated on 06/10/2011. Maintained by Guangliang Chen.